Hyperliquid fills quickstart - SQD Documentation

Pipes SDK streams Hyperliquid trade executions from the hyperliquid-fills Portal dataset. The dataset is real-time and starts at block 750,000,000.

This source covers HyperCore fills. HyperEVM uses the standard EVM Pipes SDK with the hyperliquid-mainnet or hyperliquid-testnet dataset.

The Pipes CLI currently scaffolds EVM and SVM projects only. Set up Hyperliquid and Tron pipes manually as shown here.

Create the project

Node.js 22.15 or later is required.

mkdir hyperliquid-pipe && cd hyperliquid-pipe
npm init -y
npm pkg set type=module
npm install @subsquid/pipes@beta tsx

Create index.ts:

import {
  hyperliquidFillsPortalStream,
  hyperliquidFillsQuery,
} from '@subsquid/pipes/hyperliquid'

const stream = hyperliquidFillsPortalStream({
  id: 'hyperliquid-btc-fills',
  portal: 'https://portal.sqd.dev/datasets/hyperliquid-fills',
  outputs: hyperliquidFillsQuery()
    .addFields({
      block: { number: true, timestamp: true },
      fill: { user: true, coin: true, px: true, sz: true, side: true, tid: true },
    })
    .addFillRequest({
      range: { from: 'latest' },
      request: { coin: ['BTC'] },
    }),
})

for await (const { data } of stream) {
  for (const block of data) {
    for (const fill of block.fills) {
      console.log(block.header.number, fill)
    }
  }
}

Run it:

npx tsx index.ts

The fills dataset currently exposes zero-filled block hashes. Use portal: { url, finalized: true } with persistent targets so recovery does not depend on distinguishing forks by block hash.

Next steps

Handle fills for trade-side, volume, and deduplication semantics.
Query builder reference for every field and filter.
Use the shared Postgres, ClickHouse, BigQuery, or Parquet target APIs to persist the stream.

​Create the project

​Next steps

Create the project

Next steps